congatec enables rugged AI computing with Intel Core Ultra COMs

Industrial automation, robotics, edge AI systems, and smart infrastructure require computing platforms capable of maintaining performance under extreme environmental conditions while supporting high AI workloads. In this context, congatec has introduced rugged Computer-on-Modules (COMs) based on Intel Core Ultra Series 3 processors, designed to operate reliably across an extended temperature range from -40 °C to +85 °C while delivering up to 180 TOPS of AI performance.

The modules are intended for deployment in environments exposed to temperature fluctuations, vibration, and continuous operation, where standard commercial-grade hardware may not meet reliability requirements.

Ruggedised COMs for edge AI and industrial environments
The new COM variants extend congatec’s existing portfolio by introducing industrial-grade temperature support without compromising computing performance. These modules integrate up to 16 CPU cores delivering up to 10 TOPS, alongside a dedicated neural processing unit (NPU) capable of up to 50 TOPS for low-power AI inference.

Additional AI acceleration is provided through integrated graphics with up to four Xe3 cores, enabling general-purpose GPU (GPGPU) processing. This combination allows developers to run AI inference, sensor fusion, automation control, and safety-critical workloads locally at the edge, reducing latency and dependence on cloud infrastructure.

To ensure durability, the modules can be enhanced through customisation services such as conformal coating, extended burn-in testing, and component selection tailored for harsh operating conditions.

Scalable performance across multiple form factors
The portfolio includes several module standards to support both new system designs and upgrades of existing platforms. High-performance applications requiring maximum data throughput can utilise COM-HPC modules with PCIe Gen 5 and USB4 interfaces, while compact and legacy systems can be upgraded using COM Express-based modules.

Variants such as COM-HPC Mini and Client modules target high-bandwidth applications, whereas COM Express Type 10 and Compact modules provide compatibility with existing embedded designs. Ruggedised versions with features such as screw-locked LPCAMM2 memory improve mechanical stability in vibration-prone environments.

This scalability enables system designers to select configurations ranging from energy-efficient compact systems to high-performance platforms with extensive I/O capabilities.

Software integration and workload consolidation
The modules support multiple operating systems, including Microsoft Windows 11, Windows 11 IoT Enterprise, Linux distributions, and real-time operating systems. Pre-configured software environments are available to accelerate deployment in industrial applications.

Virtualisation capabilities enable consolidation of multiple workloads—such as human-machine interface (HMI), real-time control, AI processing, and IoT gateway functions—onto a single hardware platform. This reduces system complexity and hardware footprint while improving resource utilisation.

For industrial IoT applications, software components support connectivity, remote management, and integration with cloud platforms, facilitating data exchange and system monitoring across distributed deployments.

Applications in mission-critical and outdoor systems
The extended temperature range and rugged design make these COMs suitable for mission-critical systems, including autonomous vehicles, heavy-duty machinery, and railway infrastructure. They are also applicable in outdoor edge computing scenarios such as smart city deployments, renewable energy systems, and roadside or wayside infrastructure.

In such environments, the combination of high AI performance and environmental resilience enables real-time decision-making and continuous operation under demanding conditions.

Positioning within embedded AI computing platforms
Computer-on-Modules are widely used in embedded systems to provide scalable processing capabilities while simplifying system design. Comparable solutions include modules from Advantech and Kontron, which also offer industrial-grade COMs with AI acceleration and extended temperature support.

Key selection criteria in this segment include AI processing capability (measured in TOPS), operating temperature range, power efficiency, and compatibility with existing system architectures. The integration of CPU, GPU, and NPU resources in a single module, combined with ruggedisation options, addresses the growing demand for edge AI computing in harsh environments.

By extending high-performance COMs into the industrial temperature range, congatec enables deployment of advanced AI applications in environments where reliability and continuous operation are critical.

Edited by Natania Lyngdoh, Induportals Editor — Adapted by AI.

www.congatec.com